AN ELECTROCHEMICAL REDUCTION PROCESS FOR THE RECOVERY OF COPPER POWDER FROM A REFINERY EFFLUENT STREAM

Transcription

1 AN ELECTROCHEMICAL REDUCTION PROCESS FOR THE RECOVERY OF COPPER POWDER FROM A REFINERY EFFLUENT STREAM Chandon Bezuidenhout Masters in Chemical Engineering - Hydrometallurgy University of Cape Town University of Cape Town February 214

2 The copyright of this thesis vests in the author. No quotation from it or information derived from it is to be published without full acknowledgement of the source. The thesis is to be used for private study or noncommercial research purposes only. Published by the University of Cape Town (UCT) in terms of the non-exclusive license granted to UCT by the author. University of Cape Town

3 DECLARATION I know the meaning of plagiarism and declare that all the work in the document, save for that which is properly acknowledged, is my own. C. Bezuidenhout Page i

4 EXECUTIVE SUMMARY In recent years a significant amount of research has gone into the development of a feed pretreatment process for the concentrate refined at Anglo American Platinum s Precious Metals Refiners (PMR). Such a process has the potential to significantly simplify the downstream refining process and reduce the number of unit processes required for purification. One of the considered options involves a high temperature oxidative roast process followed by a high temperature hydrochlorination process to volatilise base metal chlorides and other impurities. The resulting precious group metal (PGM) concentrate is cleaner and thus requires significantly less process steps to final product. The off-gas from hydrochlorination contains predominantly silver, copper, nickel and iron. This off-gas undergoes a quench-scrub to condense the base metal chlorides. The quench-scrub liquor undergoes a dechlorination process with sulfuric acid to precipitate AgCl(s). The filtrate from this process then undergoes an electrochemical reduction process to recover copper metal concentrate. The objectives of this study were to: i. Determine the operating conditions for an electrochemical reduction process aimed at recovering copper as a copper powder from a sulfuric acid stream containing copper, nickel and iron. ii. Develop a conceptual flowsheet for a batch electrochemical process and estimate the capital cost and operating cost. Bench-scale experimental work was conducted using two 4 L electrochemical reactors with titanium cathodes and dimensionally stable anodes. Synthetic solutions were made up to the anticipated feed concentrations of 3 g/l copper, 1 g/l nickel, 6 g/l iron and 1 mol/l sulfuric acid. Solutions were treated at high current densities ( 1 A/m 2 ) to promote the formation of copper powder. After seven hours of electrochemical reduction the electrolyte and solids were removed from the cell, filtered and washed. In order to determine the optimal operating conditions and impact of feed variability on the process, the following parameters were varied: applied current; initial acidity; initial copper concentration; initial nickel concentration; and initial iron concentration. It was found that after seven hours of electrochemical reduction near-complete reduction of Cu 2+ (aq) to Cu(s) could be achieved for all experimental runs. The Cu(s) that formed did not adhere to the cathode surface and gravitated to the bottom of the cell for subsequent recovery. After filtration and washing of the solids, various degrees of copper powder re-dissolutions were observed. During the electrochemical reduction the nickel and iron in solution did not reduce to metallic form. However, under mass transport limited conditions it was observed that ferric ions were reduced to ferrous ions and re-oxidised to ferric ions, thus reducing the current efficiency of the process. During electrochemical reduction hydrogen gas evolved at the cathode and oxygen gas evolved at the anode. For a Cu electrochemical reduction process targeting a process yield of 95% at a current density of 1 A/m 2, the required processing time was determine to be 4.5 hours with an overall current efficiency of 86.4%, cell voltage of V and specific energy consumption of kwh/kg Cu. Lower initial acidities (.5 mol/l H + ) and initial copper concentrations (3 g/l Cu) resulted in higher cell voltage, elevated electrolyte temperature and higher degrees of copper powder re-dissolutions during filtration and washing. A similar degree of copper powder re-dissolution was observed for a solution containing higher initial copper concentrations (5 g/l). Page ii

5 Feed solutions with higher acidities (1-4 mol/l H + ) and feed solutions with variation in nickel concentration (5-2 g/l) or iron concentration (1-12 g/l) had no detrimental impact on the electrochemical reduction of copper. It was found that the current applied during the electrochemical reduction process impacted on the rate of copper recovery and the specific energy consumption of the process. Based on a targeted process yield of 95% the optimum current density was found to be 1 A/m 2. A conceptual flowsheet was developed for a batch electrochemical reduction process to treat an anticipated feed solution of 22.4 kg/h Cu. The capital cost of the process was estimated at R 15.9 million and the operating cost was estimated at R 3.8 million per annum. Based on a discounted rate of 11.5% the project has a net present value (NPV) of R 26.9 million and an internal rate of return (IRR) of 37.6% over 5 years. The main recommendations from this study are to: 1. Perform larger scale piloting and demonstration of the integrated process on actual process solutions. This includes an investigation into a suitable solids handling process to convert copper powder into a saleable product and characterisation of the morphology and quality of copper solids. A detailed costing and economic evaluation study should then be done based on the pilot plant results. The main factors to consider during design of the pilot plant include: batch vs. continuous operation; optimisation of anode and cathode materials of construction; optimisation of electrode spacing; hydrodynamics of electrochemical cell; conditions for removing copper deposits; solid purity; particle size distribution; solids morphology and filtration flux. 2. Perform a detailed economic evaluation to compare the OPEX and CAPEX for the electrochemical reduction process to that of the alternative hydrometallurgical processes considered. This includes further study to determine the operating conditions and economic parameters for a suitable process to recover nickel from the copper spent electrolyte. Page iii

6 TABLE OF CONTENTS DECLARATION... I EXECUTIVE SUMMARY... II LIST OF TABLES...VII LIST OF FIGURES...VIII NOMENCLATURE...X 1 INTRODUCTION Background Problem Statement Objectives Methodology Hypothesis LITERATURE SURVEY The Recovery of Copper from Effluent Streams Commercial Processes for the Production of Copper Powder Atomised Copper Powder Electrochemical Copper Powder Hydrometallurgical Copper Powder Solid State Reduced Copper Powder Electrochemical Methods For Producing Metal Powders FUNDAMENTALS Chemistry The Effect of Current Density The Effect of Feed Copper Concentration The Effect of Feed Nickel Concentration The Effect of Feed Iron Concentration Thermodynamics The Thermodynamic Order of Reducibility The Effect of Complexation on Redox Potentials Electrochemical Process Parameters Process Yield Current Efficiency Cell Voltage Electrical Energy Consumption Electro-active Area Per Unit Volume EXPERIMENTAL Page iv

7 4.1 Experimental Design Experimental Setup Electrochemical Reactor DC Power Supply Data Acquisition System Experimental Procedure RESULTS AND DISCUSSION Electrochemical Reduction to Produce Copper Powder Aqueous Concentration Profiles Reproducibility Process Yield Current Efficiency Cell Voltage and Energy Consumption Electro-active Area and Process Kinetics Gas Evolution and Solids Formation Optimal Operating Conditions and Impact of Feed Variation Applied Current Initial Acidity Initial Copper Concentration Copper Powder Re-dissolution Summary Flowsheet Development Process Description Process Schedule Equipment Sizing Capital Cost Estimation Operating Cost Estimation Economic Evaluation CONCLUSIONS AND RECOMMENDATIONS REFERENCES APPENDIX A THERMODYNAMICS Standard Half-Reaction Potentials in Acid Solutions APPENDIX B OPERATING AND RECOVERY PROFILES Applied Current (Run 1A, 2 & 3) Initial Acidity (Run 1A, 4, 5) Initial Copper Concentration (Run 1A, 6, 7) Initial Nickel Concentration (Run 1A, 8, 9) Page v

8 Initial Iron Concentration (Run 1A, 1, 11) APPENDIX C CALCULATIONS Page vi

9 LIST OF TABLES Table 1: Properties of copper powder by various industrial processes [11,12] Table 2: Electrochemical methods for copper powder production (adapted from [11])... 1 Table 3: Experimental design conditions Table 4: Experimental design matrix Table 5: Summary of process yields and overall current efficiency for all experimental runs after seven hours of processing Table 6: Processing parameters at various current densities and targeted Cu extractions Table 7: Equipment sizes based on a design factor of 15% Table 8: Mechanical cost calculation and capital cost estimation for the major plant equipment Table 9: Estimated operating costs per annum Page vii

10 LIST OF FIGURES Figure 1: Potential feed pre-treatment for PGM concentrate to PMR Figure 2: Typical copper powders particles formed in (a) a nitrogen atomised process and (b,c) water atomised processes [11] Figure 3: Flow diagram for a water atomisation process for production of copper powder [11]... 7 Figure 4: Process route for electrochemical copper powder production [11] Figure 5: Current efficiency and apparent density for copper powder as a function of (a) copper concentration and (b) sulfuric acid concentration [13] Figure 6: Electrochemical cell for production of copper powder [11]... 9 Figure 7: Dendritic structures of copper particles after (a) electrochemical treatment and (b) subsequent thermo-mechanical treatment [11]... 1 Figure 8: Process route for hydrometallurgical copper powder production [11] Figure 9: Copper powder particles produced at a current density of 1 5 A/m 2 and increasing various copper concentrations. (a) porous-disperse, (b) cauliflower-like, (c) coral-like, (d) shrub-like and (e) stalk-stock-like [14] Figure 1: Electrochemical system for the reduction of copper in acidic solution Figure 11: Effect of current density on the nature of metal deposits [24,25] Figure 12: Schematic current-potential curves for iron competing during copper electrowinning [24-26] Figure 13: Thermodynamic order of reducibility of metal ions in standard solutions and non-standard solutions Figure 14: Species distribution diagrams for.47 M CuSO 4,.17 M NiSO 4 and.5 M Fe 2 (SO 4 ) 3 solutions at 6 C (generated in HSC [31])... 2 Figure 15: Experimental setup Figure 16: Experimental setup for a typical run Figure 17: Aqueous concentration profiles (Run 1B) Figure 18: Aqueous Fe 3+ /Fe 2+ speciation profiles (Run 1B) Figure 19: Aqueous concentration profiles to illustrate reproducibility (Run 1A and 1B) Figure 2: Aqueous Fe 3+ /Fe 2+ speciation profiles to illustrate reproducibility (Run 1A and Run 1B) Figure 21: Process yield for copper (Run 1A and 1B) Figure 22: Instantaneous current efficiency (Run 1A and 1B) Figure 23: Overall current efficiency (Run 1A and 1B) Figure 24: Cell voltage (Run 1A and 1B) Figure 25: Electrolyte temperatures (Run 1A and 1B) Figure 26: Linear relationship between cell voltage and applied current Figure 27: Calculated internal cell resistance based on Run 1-3 data Figure 28: Difference in equilibrium potential (Run 1A and 1B) Figure 29: Sum of the anode and cathode overpotentials (Run 1A and Run 1B) Page viii

11 Figure 3: Specific energy consumption versus process yield (Run 1A) Figure 31: Current limited and mass transport limited regimes for copper Figure 32: Pictures at various stages of the experimental run under base case conditions. (a) Evolution of oxygen gas at the anode under current limited conditions. (b) Initial stages of hydrogen gas evolution at the cathode during mixed current limited and mass transport limited conditions. (c) Evolution of oxygen gas at the anode and hydrogen gas at the cathode under mass transport controlled conditions. (d) Solids generated after seven hours of electrochemical reduction. (e) The cathode surface after seven hours of electrochemical reduction. (f) Solids remaining in the reactor after draining of the electrolyte Figure 33: Rate of Cu recovery (current limited regime) at various applied currents Figure 34: Impact of applied current on various process parameters Figure 35: Specific energy consumption and processing times as a function of current density Figure 36: Impact of initial acidity on various process parameters Figure 37: Copper powder produced under (a) Low Acidity and (b) High Acidity conditions Figure 38: Impact of initial copper concentration on various process parameters Figure 39: Pictures of the powders produced with (a,c) High Cu and (b,d) Low Copper runs Figure 4: Degree of copper powder re-dissolution during filtration and washing Figure 41: Conceptual flowsheet for the electrochemical recovery of copper Figure 42: Process schedule illustrating parallel processing of Cell 1, 2 and 3 for six batches Figure 43: Cumulative discounted cash flow diagram Figure 44: Operating and recovery profiles for of applied current (Runs 1A, 2 & 3) Figure 45: Operating and recovery profiles for initial acidity (Runs 1A, 4 & 5) Figure 46: Operating and recovery profiles for initial Cu concentration (Runs 1A, 6 & 7) Figure 47: Operating and recovery profiles for initial Ni concentration (Runs 1A, 8 & 9) Figure 48: Operating and recovery profiles for initial Fe concentration (Runs 1A, 1 & 11) Page ix

12 NOMENCLATURE Acronyms EC FL MMO PGM PMR PP RBMR SB TK VA Electrochemical cell Filter Mixed metal oxide Precious group metals Precious Metals Refiners Pump Rustenburg Base Metals Refiners Sample board Tank Vacuum receiver Parameters A Cell electro-active area, m 2 A s Electro-active area per unit volume, m 2 c Concentration, mol/l E CELL E e E e C E e A E e F Cell voltage, V Equilibrium potential, V Formal potential, V Equilibrium potential for the cathode, V Equilibrium potential for the anode, V Faraday constant, C/mol i Current density, A/m 2 I Current, A i L Limiting current density, A/m 2 M Mass, kg n nt n q Initial molar amount, mol Molar amount at time t, mol Number of electrons Charge passed Q Volumetric flow rate, m 3 /h R Universal gas constant, J/(mol.K) R CELL R CIRCUIT T Resistance of cell, Ω Resistance of circuit, Ω Temperature, K V R Effective reactor volume, m 3 XA Process yield of A, % Page x

13 ɸi Instantaneous current efficiency, % ɸ Overall current efficiency, % η A η C τ Anode overpotential, V Cathode overpotential, V Mean residence time, s Other e - (aq) (s) (g) Electrons Aqueous phase Solid phase Vapour phase Units A Ampere g Gram h Hour K Kelvin kg Kilogram kw Kilowatt L Litre mol Mole mm Millimetre ml Millilitre min Minute m Metre s Second V Volt Ω Ohm C Degrees Celsius Page xi

14 1 INTRODUCTION 1.1 BACKGROUND Anglo American Platinum Ltd. is the largest producer of refined platinum in the world. In 212 a total of 2.38 million ounces of platinum and 4.64 million ounces of precious group metals (PGMs) were refined at the Precious Metals Refiners (PMR) [1]. The PMR receives final concentrate from the Rustenburg Base Metal Refiners (RBMR). This concentrate is produced in a slow-cooling and magnetics separation process and typically contains 5-7% precious metals (platinum, palladium, rhodium, iridium, ruthenium, osmium, gold and silver). Typical impurities include base metals (cobalt, copper, nickel and iron), amphoteric metals (selenium, tellurium) and other impurities (lead, zinc, arsenic, antimony, bismuth, tin, silver) [2,3]. The refining processes applied at the PMR involves a series of separation and purification processes to produce high purity PGM products (>99.9%). Separation processes such as solvent extraction and distillation are used to separate specific PGMs from impurities and the remainder of PGMs. These streams are then purified through crystallisation, precipitation, hydrolysis and metal reduction [2]. In recent years a significant amount of research has gone into the development of a feed pretreatment process to remove amphoterics, base metals and other impurities from the feed concentrate refined at the PMR. The removal of these impurities from the feed concentrate has the potential to significantly improve the separation efficiencies of the overall process. It further has the potential to significantly simplify the downstream PGM refining process and reduce the number of unit processes required for purification [3]. US patent 213/ A1 [3] describes a process for the upgrading of this PGM concentrate where near-complete removal of base metals and other impurities is achieved. The patent describes a high temperature oxidative pre-treatment process followed by a high temperature hydrochlorination process to volatilise base metal chlorides and other impurities. The aim of the oxidative pre-treatment process is to convert metals to metal oxides. The metal oxide solids are then contacted with a hydrochlorination agent (at a controlled HCl/H 2 O ratio) at elevated temperatures to produce metal chlorides and water as shown in the following reaction [3]: MO + 2 HCl MCl 2 + H 2O (1) The above-mentioned reaction is thermodynamically favourable over the direct hydrochlorination of metal to form metal chlorides and hydrogen gas, and thus requires a lower stoichiometric excess of HCl(g) to volatilise an equivalent amount of base metal chlorides. Selective volatilisation of base metal chlorides and other impurities over PGM chlorides is achieved by controlling the temperature and HCl/H 2 O ratio during the hydrochlorination process [3]. During the oxidative pre-treatment process sulfur, selenium, tellurium and osmium species are volatilised to the off-gas. The species are scrubbed using a suitable off-gas scrubbing system. The off-gas from the hydrochlorination process contain predominantly HCl(g), H 2 O(g) and base metal chlorides of silver, copper, nickel and iron. This off-gas undergoes a quench-scrub process with 33% HCl to condense the base metal chlorides and recycle HCl(g). At 33% HCl the scrub solution is saturated with HCl and minimal absorption of HCl(g) is expected. This HCl(g) stream is routed to an appropriate HCl recovery process to regenerate pure HCl(g) for reuse in the hydrochlorination Page 1

15 process. The quench-scrub solution is bled to a base metal handling process for subsequent recovery value in the form of silver, copper and nickel. Figure 1 illustrates a conceptual block flow diagram for the oxidative pre-treatment process and offgas scrubbing, hydrochlorination process with off-gas quench-scrub, HCl recovery process and base metal handling process. Figure 1: Potential feed pre-treatment for PGM concentrate to PMR. Within base metal handling a process for the precipitation of silver chloride has been considered. This involves a dechlorination processes with sulfuric acid at elevated temperatures, followed by backdilution with water and filtration to recover the silver chloride concentrate. The resulting filtrate would then be concentrated and treated with a suitable hydrometallurgical process to recover copper and nickel, respectively. Hydrometallurgical process options that have been considered include the following: 1. Recycling of the base metal solution to RBMR for leaching, purification and electrowinning to produce copper cathode and nickel cathode. 2. Electrochemical reduction of copper powder followed by a suitable precipitation processes to recover nickel. 3. Electrochemical reduction copper powder followed by the electrochemical reduction of nickel powder. 4. Production of a suitable concentrate for toll-refining. Techno-economic evaluations for each of the above-mentioned process options will be done in order to develop a suitable business cases for the overall project. This study will focus on the electrochemical reduction of copper powder. 1.2 PROBLEM STATEMENT Given the relatively small stream resulting from the dechlorination process ( 72 L/h) it was considered that the sulfate solution containing approximately 3 g/l copper, 1 g/l nickel and 6 g/l iron could be rerouted to the RBMR for electrowinning of copper and nickel cathodes. However, in the event that this is impractical or not feasible, suitable processes would be required at PMR to recover both copper and nickel. Such processes would involve the production of copper and nickel concentrates that can either be recycled to Waterfall Smelter or converted into saleable products. Page 2

16 Based on the current London Metal Exchange prices the value of the respective streams are estimated to be in the order of R 15.7 million per annum of copper and R 1.2 million per annum for nickel [4,5]. Electrochemical processes with high process yields (in the order of 9-95%) were proposed where powders, rather than cathodes were produced. Due to the order of reducibility of these metals, and the different operating conditions required for the electrochemical production, two separate electrochemical processes would be considered. Firstly, the electrochemical recovery of copper powders from acidic sulfate solutions and secondly, the electrochemical reduction of nickel powder under less acidic sulfate conditions. The spent electrolyte from electrochemical recovery of nickel would then be sent for effluent treatment. The conventional copper electrowinning processes applied at RBMR typically have feed solutions in the order of 5-8 g/l Cu and the process is operated at a copper bite ( Cu) of approximately 2 g/l to produce saleable copper cathodes. However, for the proposed electrochemical reduction process for PMR it is required that effectively all copper be removed from solution in the form of a copper powder. The upfront hydrochlorination and dechlorination processes would be operated as batch, and it is thus required that the copper and nickel electrochemical processes also be operated as batch. Research and development work is required on an electrochemical process for the selective recovery of copper from a sulfuric acid solution containing predominantly copper, nickel and iron. Typical aqueous concentrations have been estimated at 3 g/l copper, 1 g/l nickel and 6 g/l iron at an acidity of 2 mol/l. The feasibility of an electrochemical process for the recovery of copper is dependent on the estimated capital and operating cost of the process to treat the required inputs. It is therefore required that experimental work be conducted to determine the operating conditions for such a process. Once the required operating conditions have been established it is required that a conceptual flowsheet be developed and estimates for the capital cost and operating cost of the process be provided. 1.3 OBJECTIVES The objectives of this study were to: i. Determine the operating conditions for an electrochemical reduction process aimed at recovering copper as a copper powder from a sulfuric acid stream containing copper, nickel and iron. ii. Develop a conceptual flowsheet for a batch electrochemical process and estimate the capital cost and operating cost. 1.4 METHODOLOGY The methodology followed for this study was as follows: i. A literature study was conducted on the recovery of copper from effluents streams and processes for the recovery of copper powders. This included hydrometallurgical processes for the recovery of copper from effluent streams, commercial processes for the recovery of copper powders and a study on the factors that impact electrochemical reduction of powders. Page 3

17 ii. iii. iv. Bench-scale experimental work was conducted using synthetic solutions made up to the anticipated feed concentrations. Parameters such as feed concentrations, feed acidity and applied current were varied to determine the impact of each on the operating conditions of the process. A conceptual flowsheet was developed based on the experimental operating conditions of the electrochemical process. Estimates for capital cost and operating cost were determined based on mechanical equipment, reagent and utility costs for the current PMR and Lang factors estimation. 1.5 HYPOTHESIS It was hypothesised that for a feed electrolyte containing 3 g/l copper, 1 g/l nickel and 6 g/l iron at an acidity of 2 mol/l, copper powder could be produced using an electrochemical reduction process at high current densities ( 1 A/m 2 ). Cu 2+ (aq) is cathodically reduced to Cu(s) that is poorly adhered to the cathode surface. The metallic copper powder can then be separated by solid-liquid separation. Page 4

18 2 LITERATURE SURVEY A literature study was conducted on the recovery of copper from effluents streams and processes for the recovery of copper powders. This included hydrometallurgical processes for the recovery of copper from effluent streams, commercial processes for the recovery of copper powders and a study on the factors that impact electrochemical reduction of powders. 2.1 THE RECOVERY OF COPPER FROM EFFLUENT STREAMS Agrawal and Sahu [6] studied processes for the recovery of copper from copper electrorefining bleed streams, solid wastes, scraps, and by-products from other processes containing significant amounts of Cu, Ni, Co. A number of pyrometallurgical- and hydrometallurgical processes were considered. Pyrometallurgical processes are generally used when the metal to be recovered is high (up to approximately 65% of the feed material), and are not suited for feeds containing less than 5% of required metal. The pyrometallurgical methods that are generally used to recover copper are the Outokumpu process, the Ausmelt smelting process and ISASMELT technology [6]. Hydrometallurgical processes for the treatment of copper ores and concentrates account for approximately 2% of global primary copper production. These processes consist of sulfate-, sulfate/chloride or chloride (and bromide) based leaching processes that are coupled with combinations of solvent extraction and electrowinning technologies. Although widely used for primary ores, hydrometallurgical processes for the recovery of copper from secondary sources have not found large application due to poor economic feasibilities [6]. Agrawal et al. [7] studied three different hydrometallurgical approaches to recover copper from copper bleed streams. These were as follows: i. Synthesis of copper sulfate powder by evaporative crystallisation; ii. Synthesis of pure copper powder by direct hydrogen reduction; and iii. Synthesis of electrolytic grade copper powder. Large and bench scale experimental work was conducted on synthetic solutions (38.42 g/l Cu g/l Ni and g/l H 2 SO 4 ) and actual copper bleed stream solutions from an Indian copper smelter (39.86 g/l Cu, 9.58 g/l Ni,.26 g/l Fe,.18 g/l Bi,.7 g/l As,.55 g/l Sb and g/l H 2 SO 4 ). In the synthesis of copper sulfate powder by evaporative crystallisation, copper bleed stream solutions were evaporated and crystallised to produce copper sulfate and mixed copper-nickel sulfate. It was found that the fractional recovery of copper crystal increases with increasing evaporation. At 55% evaporation the formation of nickel crystals became significant and a mixed copper-nickel salt was crystallised [7]. In the synthesis of pure copper powder by direct hydrogen reduction approximately 99.4% copper powder recovery could be achieved in 2 hours using an 1 L capacity titanium lined autoclave. The autoclave was operated at 453 K and 8 bar at a stirring speed of 4 rpm. The copper powder produced from actual copper bleed stream solutions had a purity of 99.4% [7]. Agrawal et al. [8] demonstrated the recovery of copper powder from copper bleed streams containing copper, nickel and other impurities (Fe, Bi, As, Sb) in sulfuric acid. The process comprised decopperisation, crystallisation, solvent extraction and electrochemical treatment. For electrochemical treatment a large scale (9 L) undivided cell with two cylindrical stainless steel cathodes and two leadantimony (6%) anodes was used. An electrolyte of g/l Cu and 9 g/l H 2 SO 4 yielded a 95% copper powder recovery after eight hours at a current density of 7 A/m 2. The morphology of the Page 5

19 copper powder was found to be dendritic in nature with a particle size of approximately 5% passing 15 µm. After annealing with H 2 gas a copper powder with purity of 99.95% was produced. Agrawal et al. [9] then went on to demonstrate the quantitative and selective recovery of nickel powder from the solvent extraction raffinate produced from the above-mentioned process. The same 9 L undivided cell was used. An electrolyte of 1 g/l Ni, 4 g/l (NH 4 ) 2 SO 4, 1 g/l H 3 BO 3 and.2 g/l thiourea yielded a 87% nickel recovery at a current density of 5 A/m 2 and feed ph of Additional runs with electrolyte solution of 2 g/l Ni yielded 99% nickel recovery at a ph of The nickel powder was dendritic and after annealing with H 2 gas, a purity of 99.89% was achieved. Based on the above-mentioned studies it has been established that relatively high purity copper powder could be produced from effluent streams containing relatively low concentrations of copper. Furthermore, given the hydrometallurgical process options evaluated in recent studies, it is possible that electrochemical production could be considered for both copper and nickel production. 2.2 COMMERCIAL PROCESSES FOR THE PRODUCTION OF COPPER POWDER Industrial processes for the production of granular copper powder typically involve an atomisation, electrochemical, hydrometallurgical, or solid state reduction process [1,11]. Table 1 provides a summary of the properties of copper powder produced by the various commercial processes [12]. Table 1: Properties of copper powder by various industrial processes [11,12]. Hydrometallurgical reduction Solid state Property Atomisation Electrochemical Copper (%) Weight loss in H 2 (%) Acid insoluble (%).5-.1 max.3 max.3 max.3 max Apparent density (g/cm 3 ) Flow (sec/5 g) none 2-35 Green strength (psi) , -25 MPa Mesh (%) From Table 1 it can be seen that similar copper powder purities are achieved for all copper powder production processes. However, it should be noted that for processes involving atomisation and solid state reduction it is required that the feed material have a purity greater than 99%. Electrolytic and hydrometallurgical processes selectively recover copper from the feed material, and are thus more suited when further refining is required. Similar weight losses in hydrogen (a measure of the oxygen content of the powder) are seen for all processes. Copper powder produced through electrochemical and hydrometallurgical processes were found to be less soluble in mineral acid, and typically had lower apparent densities [1] Atomised Copper Powder Atomisation involves contacting a molten copper stream with a gas or liquid jet-stream to form metal droplets that solidify into copper powder particles. Inert gas atomisation processes produce spherical particles, whilst water atomisation processes produce particles that vary from irregular to spherical by controlling the flow rates, water pressure, design configuration and interaction conditions between the Page 6

20 water jets and the metal stream. Scanning electron micrographs (SEM) of typical copper powder particles produced through nitrogen atomisation and water atomisation are shown in Figure 2 [1,11]. Figure 2: Typical copper powders particles formed in (a) a nitrogen atomised process and (b,c) water atomised processes [11]. Commercial atomisation processes involve superheating copper above its melting point (typically C). The stream is then directed through a nozzle at a flowrate of approximately 28 kg/min. Water pressure of 1-14 MPa is typically used to produce copper powder with a particle size of less than 15 µm. Impurities in the copper melt have a significant impact on the conductivity of the copper powder and care is taken to eliminate contamination in the smelt preparation. Control of oxygen is also required to control the shape and compactibility of the copper powder. This is done by introduction of hydrogen [9]. After atomisation the copper powder is annealed in a reducing atmosphere (temperatures of approximately 8 C), milled, classified and blended to the required particle size distribution [1]. Figure 3 illustrates a flow diagram for a water atomisation process to produce copper powders. Figure 3: Flow diagram for a water atomisation process for production of copper powder [11]. Atomised copper powder is suited for most powder metallurgy applications as a result of its high flow and good strength. It is typically used in electronic and electrical applications that require high conductivity [1] Electrochemical Copper Powder Electrochemical copper powder production involves an electrodeposition process under conditions to promote the formation of a spongy powder deposit at the cathode. Electrochemical copper powder production is done by passing current through an electrolyte with low copper ion concentration, high Page 7

21 acidity and high cathode current density. Copper anodes are used to facilitate the formation of the copper powder and additives (e.g. glucose) are added to promote uniform copper deposition. In order to meet commercial requirements control of the following parameters is required: type and quantity of additive; electrolyte temperature; electrolyte circulation rate; type and size of the anodes and cathodes; electrode spacing; and brush-down interval of cathodes [1,11]. Figure 4 illustrates a process route for electrochemical copper powder production. Figure 4: Process route for electrochemical copper powder production [11]. Electrochemical processes for the production of copper powder requires low copper content in the electrolyte to prevent adherent deposits. For electrolyte containing g/l copper the maximum current efficiency is approximately 92%. Copper concentrations greater than 33 g/l result in both lower current efficiencies and more adherent copper deposits. Apparent density and particle size of the copper powder increases with increasing copper concentration, but decreases with increasing acidity. A maximum current efficiency is achieved at a concentration of 12 g/l sulfuric acid. Beyond this acidity the current efficiency decreases and leads to passivation [11,13]. The impact of copper concentration and sulphuric acid concentration on current efficiency and apparent density of the copper powder is illustrated in Figure 5. (a) (b) Figure 5: Current efficiency and apparent density for copper powder as a function of (a) copper concentration and (b) sulfuric acid concentration [13]. Page 8

22 A high current density increases the rate of powder formation and decreases the particle size of the copper powder. For a solution containing 25 g/l copper and 12 g/l sulfuric acid the particle size fraction passing 53 µm increases from 2% to 96% with an increase in current density from 6 A/m 2 to 1 A/m 2. Increased operating temperature results in higher current efficiency, reduced cell voltage and coarser powders. However at temperatures greater than 6 C operation of the electrochemical cell becomes problematic [11,13]. Removal of the copper powder deposit from the cathode is done by mechanical brushing or shaking. The brush-down interval impacts the particle size and extended brush-down intervals generally results in coarser particles and increased apparent densities [11,13]. An electrochemical cell for the production of copper powders is illustrated in Figure 16. Each electrochemical cell consists of an arrangement of anodes and cathodes in a rubber-lined or plastic electrochemical bath. Electrolyte is introduced at the top of the cell and exits at the bottom of the cell. During electrochemical treatment dendritic copper powder is formed at the cathode and is periodically brushed. On completion of the harvesting cycle the current to the cells is stopped, spent electrolyte is drained and the copper powder slurry is removed. The copper powder slurry is filtered, washed and treated with an aqueous solution (e.g. gelatine) to prevent oxidation. This is followed by a annealing process at C under reducing conditions. The resulting powder is then milled, classified and blended to the required particle size distribution. Figure 6: Electrochemical cell for production of copper powder [11]. The particle shape of dendritic copper powder after electrochemical treatment and after subsequent treatment is given in Figure 7a and Figure 7b, respectively. After thermo-mechanical treatment the initial dendrites particle size increases by welding of smaller particles [11]. A number of commercial processes for the electrochemical production of copper powder are given in Table 2. Page 9

23 Figure 7: Dendritic structures of copper particles after (a) electrochemical treatment and (b) subsequent thermo-mechanical treatment [11]. Table 2: Electrochemical methods for copper powder production (adapted from [11]). Mode Electrolyte Current density (A/m 2 ) Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of aqueous solution Electrolysis of double layered cell Electrolysis of double layered cell Reduction of solid phase Melt electrolysis Cementation Cementation 7-35 g/l Cu 5-1 g/l H 2SO g/l Cu 163 g/l H 2SO g/l Cu g/l H 2SO g/l Cu 9 g/l NH 2HSO g/l Cu 5 g/l H 2SO g/l Cu 16 g/l H 2SO g/l Cu 16 g/l H 2SO g/l Cu 5 g/l H 2SO g/l Cu 15-2 g/l H 2SO g/l Cu, H 2SO 4.2-1% oleic acid 3.2 g/l Cu, Carbon acids in toluene g/l Na 2CO 3 (a).2-1 wt% CuCl KCl-NaCl g/l NaF H 2SO 4, ph g/l Cu H 2SO 4 Temperature ( C) (b) Process parameters and conditions Current efficiency 8-99%; particle size 2-1 µm; brush down interval 15-6 min Current efficiency 75-96%; particle size 5-6 µm; apparent density g/cm Current efficiency 6-98%; apparent density g/cm 3 ; specific surface 5-18 cm 2 /g Current efficiency 9-93%; particle size 8 µm; circulation 1L(A.h) Permanent overvoltage.3-.7 V; particle size 2-2 µm Current efficiency 1%; permanent overvoltage.3-.5 V; apparent density 2.7 g/cm Current efficiency 1%; linear current variation; particle size 2-1 µm Pulsing voltage with amplitude 6 mv; frequency 1-1 Hz; particle size 5-2 µm Current efficiency 98-99%; Pulse current; full period; pulse 1:1; particle size 5 µm; apparent density 1.5 g/cm Drum rotation 1-4 rpm; dendrite size.3-3 µm Current efficiency 93%; rotation 1-4 rpm; oleic acid changed to stearine to decrease Cu 2O yield. 1 7 Current efficiency 67-69%; raw material CuO with particle size.25 mm; copper cathode; coal anode; particle size.5 mm Two dimensional twin dendrites at C; three dimensional dendrites at 897 C Drum rotation 3 5 rpm; cementing agent is aluminium Cementing agent is zinc powder; mixing (Reynolds number = ) Copper powders produced through electrochemical processes typically exceed purities of 99.5%. Electrochemical copper powder is suited for powder metallurgy applications where high electrical and thermal conductivities are required and is typically used for electrical and electronic applications [1,11]. Page 1

24 2.2.3 Hydrometallurgical Copper Powder Hydrometallurgical processes for copper powder production typically involve leaching of copper sulphide and copper oxide ores with either sulfuric acid or ammoniacal solutions. The pregnant leach solution is filtered and the filtrate undergoes precipitation of the metal from solution. Typical precipitation processes include hydrogen reduction, cementation and electrowinning [1,11]. Figure 8 illustrates a process route for hydrometallurgical copper powder production. Figure 8: Process route for hydrometallurgical copper powder production [11]. Copper powder precipitation by means of a reduction with hydrogen, carbon monoxide or sulfur dioxide can be done with both sulfurous and ammoniacal solutions. The reduction processes are done under pressure and at elevated temperatures. The following reactions describe the precipitation of copper from solution using a reduction process [11]: 2+ Cu + H Cu + 2H 2 + (2) 2+ Cu + CO + H O Cu + CO + 2H (3) Cu + SO + 2H O Cu + SO + 4H (4) ) 2 + H2 Cu Cu ( NH NH (5) ) 2 + CO + H2O Cu + CO Cu ( NH NH (6) ( NH ) 2 SO2 + 2H2O Cu + SO4 + 2NH4 Cu H (7) Precipitation with hydrogen is considered to be the most effective. + Page 11

25 Copper powder precipitation by cementation involves the addition of a less noble metal such as zinc or iron. The cementation reactions for copper with zinc powder and iron powder are given by the following reactions [11]: Cu + Zn Zn + Cu (8) Fe + CuSO4 Cu + FeSO 4 (9) In zinc cementation feed solutions containing 5 g/l copper are treated with a stoichiometric excess of zinc powder (particle size of 1-4 µm). This is done in a well-mixed tank at a temperature of 5 C. Copper powder purities of % are achieved. After cementation the slurry is filtered, washed, stabilised and dried. For iron cementation processes scrap iron is used which result in lower copper powder purities (75-85%) [11]. Electrowinning processes for the recovery of copper from leach solutions are used for solutions containing 25 g/l copper (or greater). The electrochemical process is operated at cathode current densities of A/m 2 and an electrolyte temperature of 6 C. Once a copper concentration of 15 g/l is reached, the electrolyte is drained and the copper powder is filtered, washed, stabilised and dried. Copper purities greater than 99.9% are achieved [11]. A second electrowinning process for the recovery of copper powder involves the electrochemical regeneration of titanium in a divided cell [11]. Titanium acts as a reductant for copper based on the following reaction: Cu ( aq) + 2Ti ( aq) Cu( s) + 2Ti ( aq) (1) The above-mentioned process is operated at a temperature of 25 C with cathode current density of 5 A/m 2 and anode current density of 4 A/m 2 [11]. Hydrometallurgical copper powder is typically used in friction materials due to its fine particle sizes, low apparent density and high strength [1] Solid State Reduced Copper Powder Solid state reduction involves a process where copper oxide is reduced to copper at elevated temperatures by a reducing gas. This process yields a sintered porous cake that is subsequently ground to a powder [9]. Raw material in the form of copper mill scale, cement copper or high purity chopped scrap and atomised powder is oxidised to either cuprous oxide or cupric oxide. This is done either in a rotary kiln or fluidised bed with air at temperatures greater than 65 C. The oxidation reactions for the formation of cuprous oxide and cupric oxide are as follows [11,13]: Cu( s) + 1 O ( g) Cu O( s ) (11) Cu s) + 1 O ( g) CuO( s ) (12) ( 2 2 The oxidised metal is melted in a fuel-fired or induction furnace with hydrogen, cracked ammonia or carbon monoxide at C. The reduction reactions for copper oxide by hydrogen and carbon monoxide are given by the following reactions [11,13]: Cu O s) + H ( g) 2Cu( s) + H O( ) (13) 2 ( 2 2 g Page 12

26 Cu O s) + CO( g) 2Cu( s) + CO ( ) (14) 2 ( 2 g CuO s) + H ( g) Cu( s) + H O( ) (15) ( 2 2 g Control of impurities such as aluminium, silica, tin and lead is important to prevent formation of abrasive powders, clogging of the furnace equipment and reduction in the electrical and thermal conductivity of the copper powder. The molten copper then undergoes an atomisation process with either air or water [11]. Copper powder by solid state reduction is suited for most powder metallurgy applications as a result of its high flow and good strength. Its most importance application is for self-lubricating bronze bearings [1,11]. 2.3 ELECTROCHEMICAL METHODS FOR PRODUCING METAL POWDERS Electrochemical methods can be used to produce metal powders in a variety of purities, particle shapes, particle sizes and degrees of compressibility. By varying the process parameters for these electrochemical methods, it is possible to control the crystallisation of metal powders that are formed. The electrochemical methods that are generally employed to produce metal powders (copper, nickel, cobalt, cadmium, zinc, silver) from aqueous solutions are differentiated into two deposition classes. The first class involves direct deposition of a friable or spongy deposit that disintegrates into fine particles. The second class involves deposition of a dense, smooth, brittle layer that can be ground into a powder. Both deposition classes are achieved by controlling or manipulating the following process parameters [11]: i. Concentration of metal and ph of the electrolyte ii. Current density iii. Cathode potential iv. Temperature and rate of circulation of electrolyte v. Type and size of anode and cathode and their distance from each other vi. Type and quantity of additional agents vii. Conditions of removing deposits at the electrodes Electrochemical metal powders are mainly produced as loose dendritic deposits. However, depending on the conditions of the electrochemical process and nature of the metal, powders in the form of flakes, needles, fibrous or spongy forms are observed [11]. Orhan et al. studied the effect of electrochemical parameters on the morphology of copper powder in both an electrochemical plate cell [14] and a rotating cylinder electrode cell [15]. In both studies the effect of copper concentration, electrolyte temperature and current density were evaluated in a sulfuric acid solution. With an electrochemical plate cell it was found that the morphology of the copper powder was dependent on the copper concentration and the electrolyte temperature. Electrochemical treatment of an electrolyte containing.12 mol/l Cu and.5 mol/l H 2 SO 4 at a current density of 1 5 A/m 2 and potential of 1 mv resulted in the formation of porous and disperse copper powder. With increasing copper concentration from.12 to.63 mol/l Cu, the morphology changed from porous, disperse to cauliflower-like, coral-like, shrub-like and stalk-stock-like [14]. SEM photomicrographs for the copper powder particles obtained at the various copper concentrations are given in Figure 9. The cauliflower-like structure obtained at high current densities and high overpotentials was found to be indicative of the limiting diffusion current density where the hydrogen evolution rate is high. As the Page 13

27 hydrogen evolution rate becomes lower (with increasing copper concentration) the morphology of the powder particles change as depicted in Figure 9. Furthermore, it was found that increasing temperature from 3 C to 45 C and 6 C resulted in an increased surface area and lower apparent density of copper powder solids [14]. (a) (b) (c) (e) (f) Figure 9: Copper powder particles produced at a current density of 1 5 A/m 2 and increasing various copper concentrations. (a) porous-disperse, (b) cauliflower-like, (c) coral-like, (d) shrub-like and (e) stalk-stock-like [14]. With a rotating disk electrode cell it was found that an increase in current density or decrease in electrolyte temperature resulted in a reduction in the size of the copper powder particles, thus promoting dendritic structures. Increasing copper concentration or cathode rotation speed increased the size of the particles which promoted cauliflower-like morphology [15]. A number of authors have studied the morphology of copper powder through electrochemical processes, and its correlations with hydrogen evolution [16-22]. Nikolić et al. determined three morphology groups for copper powder in sulfate solutions based on the average current efficiency of hydrogen evolution, ɸ avg,h2 [22]. These are as follows: i. ɸ avg,h2 = -1%. Branch dendrites, independently formed cauliflower-like forms and possible formation of holes due to detached hydrogen bubbles (dish-like holes). ii. ɸ avg,h2 = 1-2%. Mixture of dish-like holes and holes constituting honeycomb-like structures with independently formed copper grain agglomerates. Mixture of holes and cauliflower-like forms. Mixture of holes and branchy dendrites. iii. ɸ avg,h2 >2%. Mixture of holes and cauliflower-like copper grain agglomerates that are formed around holes. No dendrites appear. Wang et al. [23] studied the electrochemical production of copper powder under a super gravity field in an attempt to enhance the kinetics of the process. Under the experimental conditions it was found that the current efficiency of the process increased by 2% relative to normal gravity conditions. This was attributed to disturbance of the electrode/electrolyte interface which resulted in enhanced mass transfer of Cu 2+. Lower cell voltages were recorded and it was found that particle size decreased with increasing gravitation force. The use of a super gravity force furthermore facilitated the detachment of copper powders from the electrode surface. Page 14

28 3 FUNDAMENTALS This section outlines the chemistry, thermodynamics and kinetics of the electrochemical reduction process. 3.1 CHEMISTRY The electrochemical reduction of copper in acidic solutions is achieved at the cathode. This is associated with hydrogen gas evolution at the cathode and oxygen gas evolution at the anode. The main reactions for the electrochemical process are given as follows: At the cathode: Cu 2+ ( aq) + 2e Cu( s) (16) At the anode: + 2H ( aq ) + 2e H 2 ( g ) (17) + 2H 2 O( aq ) O2 ( g ) + 4H ( aq ) + 4e (18) Under mass transport limited conditions for copper it is possible for ferric to be reduced to ferrous at the cathode, which is subsequently oxidised back to ferric at the anode. This is given by the following reactions: At the cathode: Fe 3+ At the anode: 2+ ( aq ) + e Fe ( aq ) (19) (2) Fe ( aq ) Fe ( aq ) + e The above-mentioned electrochemical system is illustrated in Figure 1. V e - ANODE CATHODE O 2(g) H 2(g) 4e - + 4H + H 2O 2H + + 2e - e - Fe 2+ Fe 3+ + e - Cu e - ELECTROLYTE Cu(s) Figure 1: Electrochemical system for the reduction of copper in acidic solution. Page 15

29 Reduction of nickel is not expected to take place due to the significantly lower reduction potential relative to that of copper (See Section 3.2.1) The Effect of Current Density The current density (i) of metal deposition affects the morphology of the deposit. As current density (or the ratio of i/i L ) increases, the mean size of the crystallites making up the deposit decreases. Thus, the rate of nucleation increases. Figure 11 indicates the typical effect of current density on the nature of metal deposits [24, 25]. Coarse powders Figure 11: Effect of current density on the nature of metal deposits [24,25]. For industrial electrowinning applications, the maximum current density that is typically applied is 3-4% of the limiting current density. Thus, in order to produce a powdery deposit, the current density should be operated in the range of 5-1% of the limiting current density The Effect of Feed Copper Concentration The copper ion concentration in the electrolyte affects the morphology and power consumption of the process. As the copper ion concentration decreases, the mean size of the crystallites making up the deposit decreases, thus forming smaller, weak deposits. Low metal concentrations in the electrolyte result in low electrical conductivity. This, in turn, causes high electrical power consumption. At low copper concentrations, the overall rate-determining step of the deposition of copper is mass transport controlled. Higher metal concentrations could result in anode passivation and excessively high metal concentrations could increase the electrolyte viscosity, which results in increased solids suspension The Effect of Feed Nickel Concentration The nickel ion concentration in the electrolyte affects the morphology and power consumption of the process for elements such as copper. Page 16

30 Low metal concentrations in the electrolyte result in low electrical conductivity. This, in turn, causes high electrical power consumption. Higher metal concentrations could result in anode passivation and excessively high metal concentrations could increase the electrolyte viscosity, which results in increased solids suspension The Effect of Feed Iron Concentration The presence of iron affects the current efficiency for copper electrowinning. As the iron concentration increases, the current efficiency generally decreases. Thus, the power consumption increases. This is mainly driven by the reduction of ferric ions in solution or produced by oxidation of iron at the anode. The ferric ions are reduced at their mass transport controlled rates at the potential of the cathode. Figure 12 schematically depicts current-potential curves for the cathodic processes for copper electrowinning in the presence of iron and oxygen [24-26]. Figure 12: Schematic current-potential curves for iron competing during copper electrowinning [24-26]. 3.2 THERMODYNAMICS The following parameters were evaluated in order to illustrate the thermodynamic properties of the electrochemical process: i. The thermodynamic order of reducibility for metal ions in solution. ii. The effect of complexation on redox potentials for metal ions in solution The Thermodynamic Order of Reducibility The equilibrium potential for an electrode is calculated using the Nernst equation: E e = E e + RT nf c ln c 2 M M + (21) Page 17

31 where E e is the equilibrium potential (V), E e is the formal potential for the M 2+ /M couple in the electrolyte (V), R is the universal gas constant (J/(mol.K)), T is the temperature (K) and c is concentration (mol/l). A selection of standard half-reaction potentials in acid solutions is given Appendix A [27]. Based on these values, the thermodynamic order of reducibility for metal ions in standard solutions can be determined. Using the Nernst equation the order of reducibility can be determined for non-standard solutions. The order of reducibility for both standard and non-standard solutions is given in Figure 13. This was calculated using the standard half-reaction potentials (Appendix A) and the Nernst equation. Nonstandard solutions were based on an acidic sulfate solution at a temperature of 6 C with 3 g/l copper, 1 g/l nickel, 9 g/l iron with 2 mg/l impurities (gold, iridium, silver and rhodium). Standard Reduction Potential (V) Order of Reducibility of Metal Ions in Standard Solutions Au Ir Ag Rh Pt Pd Cu H2 Pb Sn Ni Co Cd Fe Zn Cr Mn Al Mg Reduction Potential (V) Order of Reducibility of Metal Ions in Non-standard Solutions at 6 C Au Ir Ag Rh Pt Pd Cu H2 Pb Sn Ni Co Cd Fe Zn Cr Mn Al Mg -3. Relative to SHE -3 3 g/l Cu + 1 g/l Ni + 6 g/l Fe + 2 mg/l other Figure 13: Thermodynamic order of reducibility of metal ions in standard solutions and nonstandard solutions. From Figure 13 it can be seen that the reduction potentials of gold, iridium, silver, rhodium, platinum and palladium are greater than copper for both standard and non-standard solutions. Thus, based on thermodynamics it is expected that these metals will co-deposit with copper during the electrochemical reduction. Conversely metals such as nickel and iron will not co-deposit The Effect of Complexation on Redox Potentials The formation of complexes between a metal ion and a ligand can have significant effects on the reduction potential for electrochemical couples involving the metal ion. In general, complexation reduces the activity of the free metal ion. Thus a more negative potential is required to reduce the metal ion. In order to quantify the effect of complexation under different conditions, the speciation of a metal-containing aqueous solution is evaluated in the form of species distribution diagrams. Metals in aqueous solutions are generally present in the form of soluble species such as free ions, neutral and charged complexes. The speciation of these aqueous solutions is dependent on the total concentration of each metal, the solution acidity, redox potential, temperature and pressure. Thermodynamic modelling of the chemical equilibria for the CuSO 4 -H 2 SO 4 -H 2 O system [28] in the concentration range of -2 g/l H 2 SO 4, -5 g/l Cu(II) and temperature range of 15-7 C, have indicated that a number of species can be formed at various ph and temperature values. The - predominant species that are present in solution are as follows: HSO 4 H +, SO 2-4, Cu 2+ and CuSO 4 (aq). At ph values lower than 1 it was found that sulfuric acid speciates as bisulfate ion (HSO - 4 ) and hydrogen ion (H + ). Page 18

32 The following aqueous equilibrium reactions were considered for the CuSO 4 -H 2 SO 4 -H 2 O system: Cu 2+ + SO 2 4 CuSO 4 ( aq) (22) SO 4 HSO4 H (23) Similar results have been obtained for the NiSO 4 -H 2 SO 4 -H 2 O [29] system with the following aqueous equilibrium reactions: Ni 2+ + SO 2 4 NiSO 4 ( aq) (24) SO 4 HSO4 H (23) Thermodynamic modelling of the chemical equilibria for the Fe(II)-Fe(III)-H 2 SO 4 -H 2 O system in concentrated solutions (2.2 M H 2 SO 4 and 1.3 M Fe) at temperatures of 25 C and 5 C were also studied [3]. In aqueous sulfuric acid solutions iron is distributed as Fe(II) and Fe(III) species in the form of free ions (Fe 2+, Fe 3+ ) or complex compounds (FeSO 4, FeSO 4 +, Fe(SO 4 ) 2 -.The concentration of the these species are highly dependent on the solution concentration and temperature. It was found that the predominant species under the indicated conditions are as follows: HSO 4, H +, Fe 2+ and FeH(SO 4 ) 2. Due to the high concentration of bisulfate ion (HSO 4 - ) in solution, it was proposed that aqueous solutions under these conditions demonstrate higher buffer effects. The following aqueous equilibrium reactions were considered for the Fe(II)-Fe(III)-H 2 SO 4 -H 2 O system: 2+ 2 Fe + SO4 = FeSO4 o (25) SO4 = FeSO4 Fe (26) SO4 = Fe( SO4 ) 2 Fe (27) SO 4 = HSO4 H (23) Species distribution diagrams for.47 M CuSO 4,.17 M NiSO 4 and.5 M Fe 2 (SO 4 ) 3 solutions at 6 C with increasing acidity are illustrated in Figure 14 [31]. Page 19

33 Figure 14: Species distribution diagrams for.47 M CuSO 4,.17 M NiSO 4 and.5 M Fe 2 (SO 4 ) 3 solutions at 6 C (generated in HSC [31]). From Figure 14 it can be seen that the metal ions for the considered solution (3 g/l Cu, 1g/L Ni and 6 g/l Fe) are all present as free metal ions (Cu 2+, Ni 2+, Fe 3+ and Fe 2+ ) when the sulfuric acid concentration is in excess of.5 mol/l. Thus for a solution containing 3 g/l copper, 1 g/l nickel, 6 g/l iron and 1 mol/l sulfuric acid there will be no complexes between the metal ions and complexing ligand (SO 4 2- ) and thus no impact on the reduction potential for electrochemical couples involving the metal ions. 3.3 ELECTROCHEMICAL PROCESS PARAMETERS The following section outlines electrochemical process parameters that were used to evaluate the electrochemical process [32] Process Yield The process yield, X, for a metal A in an electrochemical batch process is given as: X A n nt = (28) n where X A is the process yield of A, n is the initial molar amount of A (mole), and n t is the molar amount of A at time t (mole). Page 2

34 3.3.2 Current Efficiency The current efficiency of an electrochemical process is given as the ratio of electrical charge used to form the desired product over the total charge passed during the electrochemical process. This is given by the following equation based on Faraday s Law: n nf φ = (29) q where n is the molar amount of starting material (mole), n is the number of electrons involved in the electrode reaction, F is the Faraday constant ( C/mol), and q is charge passed (C.s). Instantaneous current efficiency (ɸ i ) is based on the charge passed for a specific time interval, whilst overall current efficiency (ɸ) is calculated based on the cumulative charge passed from time to time t Cell Voltage Cell voltage for an electrochemical cell is given as follows: E CELL C e A e = E E η η IR IR (3) C A CELL CIRCUIT where E e A and E e C relates to the equilibrium potential for the anode an cathode reactions (V), η relates to the anode and cathode overpotential (V), I is the current applied (A); and R CELL and R CIRCUIT are the overall resistance of the cell and circuit (Ω), respectively Electrical Energy Consumption Specific energy consumption of an electrochemical process relates to the electrical energy consumed per mass of product produced, and is given as follows: Specific energy consumption = nfe CELL φm (kwh/kg) (31) where n is the number of electrons involved in the electrode reaction, F is the Faraday constant ( C/mol), E CELL is the cell voltage (V), and ɸ is the process efficiency (%) and M is the mass of product produced (kg) Electro-active Area Per Unit Volume The electro-active area per unit volume, A s, for an electrochemical cell is given as: A A s = (32) V R where A relates to the cell electro-active area (m 2 ) and V R the effective reactor volume (m 3 ). Page 21

35 4 EXPERIMENTAL This section outlines the experimental design, experimental setup and experimental procedure used in this study. 4.1 EXPERIMENTAL DESIGN Based on the literature survey it was found that a number of process parameters affect the deposition of metal powders. For the purposes of this study only metal concentration, acidity and current densities were controlled, whilst potential difference and temperature were measured. It is considered that these parameters have a large impact on the design and economics of the process. Factors such as type and size of anode or cathode, electrode distance, additional agents and removal of deposits at the electrode were considered to have a smaller impact and were not varied in this study. Based on the literature survey and fundamental of electrochemical reduction of copper, it was found that the current density should be equal or greater than 7 A/m 2 to produce granular powder. Furthermore, in order to prevent complexation of metal ions the sulfuric acid concentration should be in excess of.5 mol/ L. The experimental design conditions and experimental design matrix are given in Table 3 and Table 4, respectively. The base case experimental design conditions were 3 g/l copper, 1 g/l nickel, 6 g/l iron, 2 mol/l H + and a current density of 1 A/m 2. Low and high conditions were identified for each parameter and incorporated into the experimental design matrix. Two base case runs (Run 1 A and B) were conducted to illustrate reproducibility of the experiment. This was followed by ten runs where the low and high conditions for each of the parameters were tested. Table 3: Experimental design conditions. Parameter Units Low Base High Current density A/m Feed acidity mol/l Feed Cu concentration g/l Feed Ni concentration g/l Feed Fe concentration g/l Table 4: Experimental design matrix. Run Description Current density (A/m 2 ) Acidity (mol/l) Feed Cu (g/l) Feed Ni (g/l) Feed Fe (g/l) Run 1A Base Case A Run 1B Base Case B Run 2 High Current Run 3 Low Current Run 4 High Acidity Run 5 Low Acidity Run 6 High Cu Run 7 Low Ni Run 8 High Ni Run 9 Low Ni Run 1 High Fe Run 11 Low Fe Page 22

36 4.2 EXPERIMENTAL SETUP The experimental setup comprised an electrochemical reactor, direct current (DC) power supply, data acquisition system and computer to monitor experimental profiles (See Figure 15). Figure 15: Experimental setup Electrochemical Reactor The electrochemical reactor was constructed of polypropylene. It was assembled as two 4 L platetype cells (Cell A and Cell B) to allow for two batch runs to be conducted simultaneously. The cathodes were made out of titanium sheets (225 x 175 mm) and were blanked off to an available surface area of.24 m 2 (only one side). The anodes were made out of titanium sheets (225 x 175 mm) that have been treated with a propriety coating of tantalum oxide and iridium oxide (dimensionally stable anodes) and had an available surface area of.3 m 2. The use of dimensionally stable anodes for copper electrowinning generally provides greater current efficiencies and lower specific energy consumptions than conventional lead anodes [33]. The electrodes were spaced 84 mm apart. The positive and negative power supply connections were bolted onto the header bars of the respective anode and cathode. In order to facilitate the simultaneous operation of two cells, the cathode for Cell A was connected to the anode of Cell B (Figure 15). Thermocouples in glass tubes were submerged in the electrolyte of each cell to measure the electrolyte temperatures. Voltage clamps were connected across the cathodes and anodes of each cell to measure the cell voltage. In order to measure current a shunt was installed in the power supply line to the electrochemical reactor DC Power Supply The DC power supply that was used (Agilent Technologies N5743A) was capable providing a DC power supply within a voltage range of V and a current range of -6 A. The unit was powerfactor corrected and had functionality to operate under constant current (galvanostatic) and constant voltage (potentiostatic) operation. All experimental runs were conducted under galvanostatic conditions. Page 23

37 4.2.3 Data Acquisition System The data acquisition system consisted of the Agilent Technologies LXI 34972A data acquisition unit with multiplex input that was connected to a computer. Agilent BenchLink Data Logger 3 was used to log the data at predetermined intervals. For the testwork, the following parameters were logged: i. Electrolyte A temperature ii. Electrolyte B temperature iii. Cell voltage A iv. Cell voltage B v. Current 4.3 EXPERIMENTAL PROCEDURE Synthetic solutions were made up to the required conditions for each run using demineralised water, sulfuric acid (H 2 SO 4 ), copper sulfate (CuSO 4.5H 2 O), nickel sulfate (NiSO 4.6H 2 O) and ferric sulfate (Fe 2 (SO 4 ) 3 ). Four litres of feed solution was added to each electrochemical cell and current was applied for a duration of seven hours. Two experimental runs were conducted in parallel to demonstrate reproducibility for the base case runs. Thereafter the cells were operated independently. Samples of the electrolyte were taken at the following intervals (in minutes):, 6, 12, 18, 24, 3, 36, and 42. Samples were filtered via a syringe filter. After seven hours the current was stopped, the cell drained, the slurry filtered via vacuum filtration and the solids washed with demineralised water. The filtrate was sampled and all samples were analysed for total copper, nickel and iron concentrations by ICP-OES, free acid concentration by potentiometric titration and ferrous iron concentration by potentiometric titration. Figure 16 illustrates the experimental setup for two runs. Figure 16: Experimental setup for a typical run. Experimental work was carried out at Anglo American s Technical Solutions Research in Johannesburg, South Africa. All runs were conducted in a walk-in fume cupboard at room temperature. Page 24

38 5 RESULTS AND DISCUSSION 5.1 ELECTROCHEMICAL REDUCTION TO PRODUCE COPPER POWDER In order to characterise the electrochemical reduction of copper to copper powder under the experimental conditions, the following factors were evaluated: i. Aqueous concentration profiles of copper, together with nickel and iron ii. Reproducibility of the experimental method and analysis iii. Process yield for copper iv. Current efficiency of the process v. Cell voltage and energy consumption vi. Electro-active area per unit volume and process kinetics vii. Gas evolution and solids production Aqueous Concentration Profiles The electrochemical reduction of a species is largely dependent on the thermodynamics and kinetics of the process. For a solution that contains a number of different aqueous species, the order of reducibility of species can be estimated in theory by its thermodynamic properties. However, the kinetics of the reactions taking place control what is practically probable. In order to test the hypothesis of selective recovery of copper over nickel and iron in acidic aqueous sulfate solutions, an acidic sulfate solution containing 3 g/l Cu(aq), 1 g/l Ni(aq) and 5 g/l Fe(III)(aq) was considered. The thermodynamic order of reducibility based on the Nernst equation for this solution at 25 C is as follows: Fe 3+ (aq) > Cu 2+ (aq) > H + (aq) > Ni 2+ (aq) > Fe 2+ (aq). It is thus anticipated that ferric (Fe 3+ ) ions will be reduced to ferrous (Fe 2+ ) ions. This is followed by Cu 2+ (aq) reduction to Cu(s). Once most of the copper has been removed from solution it is expected that H 2 (g) will form. Thereafter Ni 2+ (aq) and subsequently Fe 2+ (aq) will be reduced to its solid components. Due to the reversible nature of the Fe 3+ /Fe 2+ reaction it is anticipated that once ferric is reduced at the cathode, some of it might oxidise at the anode, causing a mixture of ferric and ferrous species to be present in the electrolyte. The aqueous metal concentration profiles of copper, nickel and iron under base case conditions (Run 1B) are given in Figure 17. The ratio of ferric to ferrous ions for the same run is given in Figure 18. From Figure 17 it can be seen that the Cu 2+ (aq) concentration first decreased linearly with time, followed by an inverse exponential decrease, The nickel and total iron in solution remained constant under the experimental conditions. This illustrates that Cu 2+ (aq) was reduced to Cu(s) under the experimental conditions, without the reduction of Ni 2+ (aq) to Ni(s) or Fe 2+ (aq) to Fe(s). It furthermore illustrates that the electrochemical recovery of copper under the base case conditions was governed by two operating regimes: current limited (linear range) and mass transport limited (exponential range). Figure 17 illustrates a linear increase and gradual flattening of the H + (aq) concentration profile. This corresponds with the inverse of the Cu 2+ (aq) concentration profile. The initial linear increase in the H + (aq) concentration was a result of the water oxidation reaction that occurred at the anode at the corresponding rate. The gradual flattening of the H + (aq) concentration profile indicates evolution of H 2 (g) by the reduction reaction of H + (aq). Similar aqueous concentration profiles were seen for all the experimental runs conducted (Run 1A-11). See Appendix B. Page 25

39 Concentration (mg/l) Acidity (g/l) 5.5. Time current applied (min) Cu2+(aq) Ni2+(aq) Fe total (aq) H+(aq) Figure 17: Aqueous concentration profiles (Run 1B). 7 6 Concentration (mg/l) Time current applied (min) Fe3+(aq) Fe2+(aq) Fe total(aq) Figure 18: Aqueous Fe 3+ /Fe 2+ speciation profiles (Run 1B). Page 26

40 Figure 18 illustrates the ferric-, ferrous- and total aqueous iron concentration profiles for Run 1B. It can be seen that as the Fe 3+ (aq) concentration decreased, the Fe 2+ (aq) concentration increased to maintain a constant total aqueous iron concentration ( 5.5 g/l). Furthermore, it can be seen that the reduction of Fe 3+ (aq) to Fe 2+ (aq) occurred at two different rates. From minutes to approximately 3 minutes the rate of Fe 3+ (aq) reduction was slow. After 3 minutes, the rate of Fe 3+ (aq) reduction increased. This corresponds to the point where Cu 2+ (aq) is depleted and H 2 (g) evolution is dominant. The increased rate of Fe 3+ (aq) reduction after 3 minutes is attributed to copper becoming mas transport limited Reproducibility In order to illustrate reproducibility of the experimental method, the base case runs (Run 1A and Run 1B) were conducted using the same feed solution under the same operating conditions. Aqueous concentration profiles for Run 1A and Run 1B are compared in Figure 19 and Figure Concentration (mg/l) Acidity (g/l) 5.5. Duration (min) Run 1A Cu (mg/l) Run 1A Ni (mg/l) Run 1A Fe (mg/l) Run 1B Cu (mg/l) Run 1B Ni (mg/l) Run 1B Fe (mg/l) Run 1A H+ (g/l) Run 1B H+ (g/l) Figure 19: Aqueous concentration profiles to illustrate reproducibility (Run 1A and 1B). Page 27

41 7 6 Concentration (mg/l) Duration (min) Run 1A Fe3+ (aq) Run 1A Fe2+ (aq) Run 1A Fe total Run 1B Fe3+ (aq) Run 1B Fe2+ (aq) Run 1B Fe total Figure 2: Aqueous Fe 3+ /Fe 2+ speciation profiles to illustrate reproducibility (Run 1A and Run 1B). Figure 19 and Figure 2 illustrate very good reproducibility for the experimental method and analysis followed. The copper concentration profile and electrolyte acidity measured for Run 1A correlated well with that of Run 1B (correlation coefficients of 99.99% and 99.97%, respectively). The nickel and iron concentrations for Run 1A were found to be statistically similar to that of Run 1B at a confidence level of 92.43% and 97.36%, respectively. The ferrous ion concentration profile for Run 1A correlated well with that of Run 1B (correlation coefficient of 99.36%) Process Yield The process yield for copper (X Cu ) under base case conditions (Run 1A and Run 1B) are given in Figure 21. From Figure 21 it can be seen that a process yield of greater than 99.99% could be achieved after seven hours under the base case conditions. Process yield for copper increased linearly with time in the current limited regime as per Faraday s law. Once the mass transport limited regime was approached; the rate of increase was slower. For a process yield of 95% it was required that current be applied to the electrolyte for the duration of approximately 4.5 hours. After electrochemical reduction for seven hours and subsequent washing and filtration it was found that the Cu 2+ (aq) concentration in the filtrate increased slightly, which resulted in a reduction in process yield. For Run 1A and Run 1B the process yields decreased from >99.99% to 99.95% and 99.93% respectively after filtration. This is attributed to the formation of very fine copper powder under mass transport limited conditions, and subsequent re-dissolution. Page 28

42 11% 1% 9% 9.57% 98.9% 99.94% 1.% 8% 73.3% Process Yield, X Cu 7% 6% 5% 4% 3% 26.12% 52.49% 2% 1% % Time current applied (min) Run 1A Run 1B Figure 21: Process yield for copper (Run 1A and 1B) Current Efficiency The instantaneous current efficiency (ɸ i ) and overall current efficiency (ɸ) of the electrochemical reduction of copper are given in Figure 22 and Figure 23 for the base case runs. It can be seen that the instantaneous current efficiency decreased from 18% to % during the course of the experimental run, whilst the overall current efficiency was in the range of 18% to 58%. As the experimental run proceeded to the point where copper was completely removed from solution (process yield of 1%), the instantaneous current efficiency and overall current efficiency decreased to % and 58%, respectively. Current efficiencies greater than 1% is attributed to non-ideal mixing of the electrolyte in the absence of hydrogen evolution. It is also considered that the formation of a Cu + intermediate could impact the current efficiency of the process [34]. The corresponding reduction reaction is given as follows: Cu + + e Cu (33) Page 29

43 Instantaneous Current Efficiency (%) 12% 1% 8% 6% 4% 2% % 18% 12% 92% 7% 32% 9% % Time current applied (min) Run 1A Run 1B Figure 22: Instantaneous current efficiency (Run 1A and 1B). 12% 1% 18% 15% 1% 92% Overall Current Efficiency (%) 8% 6% 4% 2% 8% 68% 58% % Time current applied (min) Run 1A Run 1B Figure 23: Overall current efficiency (Run 1A and 1B). Page 3

44 5.1.5 Cell Voltage and Energy Consumption The electrical energy consumption of an electrochemical process is closely related to its cell voltage. In order to minimise energy consumption of a specific process, it is preferable to select process conditions where the cell voltage is minimised. Cell voltage and electrolyte temperature profiles for the base case runs are given in Figure 24 and Figure Cell Voltage (V) Time current applied (min) Run 1A Run 1B Figure 24: Cell voltage (Run 1A and 1B). 7 6 Temperature ( C) Time current applied (min) Run 1A Run 1B Figure 25: Electrolyte temperatures (Run 1A and 1B). Page 31

45 Cell voltage measurements under base case conditions (Figure 24) decreased from 5. V to 3.2 V over a period of seven hours. This decrease in cell voltage is attributed to a change in the electrolyte temperature from 27 C to 55 C (Figure 25) due to current passing between the electrodes. The increase in electrolyte temperature increases the conductivity of the electrolyte, which, as a result of its inverse proportionality, decreases the internal cell resistance (R CELL ). The internal cell resistance for a given electrochemical cell can be determined by evaluating the linear relationship that exists between cell voltage (y-axis) and applied current (x-axis). This relationship is given as follows by re-arranging Equation 3: C A ( E e E e ) E = R I + η (34) CELL CELL y = m. x + c where the internal cell resistance is determined from the slope of the line (m); and the sum of the anode and cathode equilibrium potentials and overpotentials by the y-intercept of the line (c). Cell voltage data for three runs at different applied currents (Run 1A, 2 and 3) was used in the analysis to determine the required parameters. The linear relationship between cell voltage and applied current (Equation 34) is given in Figure 26 at three different processing times. Cell Voltage (V) y =.1178x R² =.9994 y =.42x R² =.9953 y =.257x R² = Applied Current (A) min 6 min 42 min Figure 26: Linear relationship between cell voltage and applied current. Figure 26 illustrates a decrease in the cell resistance (slope of lines) from.1178 to.257 Ω with increasing processing time. In order to illustrate the cell resistance of the process over the full processing time, the cell resistance was calculated at 1 second intervals. This is illustrated in Figure 27. Page 32

46 R CELL (Ω) Time current applied (min) Figure 27: Calculated internal cell resistance based on Run 1-3 data. From Figure 27 it can be seen that the internal cell resistance decreases from.12 to.3 Ω. This supports the notion of a decrease in cell voltage as a result of increased temperature and increased conductivity of the electrolyte. The anode and cathode equilibrium potentials can be calculated using the Nernst equation, provided the reactions that occur at the anode and cathode; electrolyte temperature; and the molar concentrations of the relevant species are known. For the purposes of this analysis the reactions are taken as follows: 2 At the cathode: Cu + 2e Cu( s) + E =.34 V (16) At the anode: + + 2H O O2( g) + 4H 4e 2 E = V (18) Thus, the cathode and anode equilibrium potential for the considered system can be expressed as follows: RT 1 E C = + e.34 ln 2 (35) 2 F [ Cu + ] + 4 ( ] ) RT Ee A = ln [ H (36) 4F The difference in anode and cathode equilibrium potential under base case conditions (Run 1A and Run 1B) was calculated using Equation 35 and Equation 36. This is illustrated in Figure 28 as a function of time. From Figure 28 it can be seen that the C E - e A E is in the order of 1.6 V and e increases to 1.73 V with increase in temperature and decrease in Cu 2+ (aq) concentration. Page 33

47 E e C - Ee A (V) Time current applied (min) Run 1A Run 1B Figure 28: Difference in equilibrium potential (Run 1A and 1B). Given the difference in equilibrium potential, the difference in overpotential can be estimated using the y-intercept (c) of the linear correlation (Equation 34). The sum of the anodic and cathodic overpotentials over time is given in Figure 29. This is in the order of V ŋ A + ŋ C (V) Time current applied (min) Figure 29: Sum of the anode and cathode overpotentials (Run 1A and Run 1B). Page 34

48 Based on the cell voltage and overall current efficiency of the electrochemical process, the specific energy consumption was determined to be in the range of kwh/kg Cu. Figure 3 illustrates the specific energy consumption of the process relative to its process yield under the base case conditions % Specfic Energy Consumption (kwh/kg Cu) % 8% 6% 4% 2% % Time current applied (min) Process Yield (%) Spec. Energy Consumption Process Yield Figure 3: Specific energy consumption versus process yield (Run 1A). From Figure 3 it can be seen that for an electrochemical process targeting a process yield of 95% the overall specific energy consumption is kwh/kg Cu Electro-active Area and Process Kinetics The electro-active area of the electrochemical cell was determined to be 6. m 2 /m 3. This is based on bench-scale cathodes area of.24 m 2 and electrolyte volume of.4 m 3. In order to illustrate the current limited and mass transport limited regimes for copper over the entire experimental range, the recovery profiles for Run 1A (Base Case), Run 6 (High Cu) and Run 7 (Low Cu) were combined and off-set for processing time. The combined data set is given in Figure 31. Page 35

49 [Cu], (mol/l) y = -.2x R² = y = 11557e -.35x R² = Time current applied (min) Linear (Current Limited) Expon. (Mass Transport Limited) Figure 31: Current limited and mass transport limited regimes for copper. Figure 31 illustrates a linear current limited regime where copper is electrochemically reduced at rate of 2.3x1-3 mol.l -1.min -1 up to a concentration of approximately.125 mol/l copper. Below.125 mol/l copper, the process undergoes a mixed zone before becoming mass transport limited below a concentration of approximately.5 mol/l copper Gas Evolution and Solids Formation In order to illustrate the evolution of gases and the formation of solids, pictures were taken during various stages of the experimental runs. This is illustrated in Figure 32. The evolution of oxygen gas at the anode under current limited conditions is illustrated in Figure 32a. During electrolysis under these conditions small oxygen bubbles are formed on the anode surface and form streaky patterns in the electrolyte. No gas evolution is observed at the cathode surface. During electrolysis the electrolyte becomes milky as a result of the oxygen bubbles and changes colour from blue to green as copper is removed from solution. The streaky patterns in the electrolyte indicate poor mixing under these operating conditions. The initial stages of hydrogen gas evolution at the cathode between the current limited and mass transport limited operating regimes are illustrated in Figure 32b. Large hydrogen bubbles are observed at the cathode surface and small oxygen bubbles are observed at the cathode. The evolution of oxygen gas at the anode and hydrogen gas at the cathode under mass transport controlled conditions are illustrated in Figure 32c. Small oxygen bubbles are observed at the anode with larger hydrogen bubbles at the cathode. Under these conditions the patterns formed by oxygen gas evolution at the anode are less streaky, thus indicating better mixing of the electrolyte. Page 36

50 (a) (b) (c) (d) (e) (f) Figure 32: Pictures at various stages of the experimental run under base case conditions. (a) Evolution of oxygen gas at the anode under current limited conditions. (b) Initial stages of hydrogen gas evolution at the cathode during mixed current limited and mass transport limited conditions. (c) Evolution of oxygen gas at the anode and hydrogen gas at the cathode under mass transport controlled conditions. (d) Solids generated after seven hours of electrochemical reduction. (e) The cathode surface after seven hours of electrochemical reduction. (f) Solids remaining in the reactor after draining of the electrolyte. The solids generated after seven hours of electrochemical reduction are illustrated Figure 32d. Once the current stopped, the electrolyte became clear. It was observed that the majority of the solids were present at the bottom of the electrochemical reactor. A small amount of solids were observed on the Page 37

51 cathode surface. However, these solids were loosely adhered to the cathode surface and detached on removal on cathode from the reactor. The cathode surface after seven hours of electrochemical reduction is illustrated in Figure 32e. The cathode surface contained some copper solids, but no electroplating of copper on the surface was observed. The copper solids could be removed by washing the surface with water and the cathode could be reused in subsequent experimental runs. The solids produced after seven hours of electrochemical reduction and draining of the filtrate is illustrated in Figure 32f. The bulk of the solids were granular. A single metal film was observed. It was considered that this was part of the solids that loosely adhered to the copper cathode surface and subsequently detached on removal of the cathode. Page 38

52 5.2 OPTIMAL OPERATING CONDITIONS AND IMPACT OF FEED VARIATION In order to determine the optimum operating regime and the impact that feed variation could have on the overall process, variation of the following parameters were considered: i. Applied current ii. Initial acidity iii. Initial Cu concentration iv. Initial Ni concentration v. Initial Fe concentration A summary of the process yield (after electrochemical reduction and after filtration) and overall current efficiencies for all experimental runs (Run 1A-Run 11) is shown in Table 5. Table 5: Summary of process yields and overall current efficiency for all experimental runs after seven hours of processing. Run Description Process yield after reduction (%) Process yield after filtration (%) Overall current efficiency (%) Run 1A Base Case A 99.98% 99.53% 59.58% Run 1B Base Case B 1.% 99.35% 59.59% Run 2 High Current 1.% 99.89% 39.51% Run 3 Low Current 99.33% 98.8% 77.91% Run 4 High Acidity 99.99% 99.87% 61.34% Run 5 Low Acidity 99.85% 9.3% 63.68% Run 6 High Cu 94.67% 9.11% 12.76% Run 7 Low Cu 99.51% 93.89% 6.5% Run 8 High Ni 1.% 99.71% 59.29% Run 9 Low Ni 99.99% 99.63% 61.62% Run 1 High Fe 99.95% 99.36% 62.4% Run 11 Low Fe 1.% 99.87% 63.18% From Table 5 it can be seen that the process yield after electrochemical reduction for all experimental runs varied in the range of %. The low process yield for High Cu (94.67%) is attributed to insufficient processing time for complete copper reduction. Under experimental conditions where sufficient processing time was allowed, the process yields after electrochemical reduction were in the range of %. The corresponding process yields after filtration and washing of the electrolyte were in the range of %. Overall current efficiencies varied from 6.5% (Low Cu) to 12.76% (High Cu). Overall current efficiencies for runs with higher and lower acidities, nickel concentrations and iron concentrations (Runs 4, 5 and 8-11) were comparable to that of Base Case A and Base Case B. Overall current efficiencies were lower than Base Case when higher current was applied (Run 2), and higher than Base Case when low current was applied (Run 3). This is attributed to a change in the rate of copper reduction under high current conditions the rate of copper reduction increases and under low current conditions the rate decreases. The operating and recovery profiles for all experimental runs are given in Appendix B. Page 39

53 5.2.1 Applied Current Based on the operating and recovery profiles for Runs 1A (Base Case), 2 (High Current) and 3 (Low Current) (Appendix B) it can be seen that the rate of copper recovery, electrolyte temperature, cell voltage and acidity increased with increasing applied current. Nickel and iron recovery profiles remained unchanged. Figure 33 illustrates the rate of copper recovery in the current limited regime at various applied currents. Rates of copper recovery under current limited conditions [Cu] (mol/l) y = -.15x R² = y = -.25x R² =.9935 y = -.18x R² = A/m2 1 A/m2 15 A/m2 Figure 33: Rate of Cu recovery (current limited regime) at various applied currents. From Figure 33 it can be seen that the rate of Cu recovery increases with increasing applied current. A 5% increase in current density (from 1 A/m 2 to 1 5 A/m 2 ) resulted in a 43% increase the rate of Cu recovery (from 1.77 x 1-3 to 2.54 x 1-3 mol/(l.min)). Electrolyte temperature and cell voltage were respectively 13 C and.38 V higher than Base Case after 3 hours of processing. A 25% decrease in current density (from 1 A/m 2 to 75 A/m 2 ) resulted in a 15% decrease in the rate of Cu recovery (from 1.77 x 1-3 to 1.49 x 1-3 mol/(l.min)). Electrolyte temperature and cell voltage were respectively 8 C and.6 V lower than Base Case after 3 hours of processing. Figure 34 illustrates the impact of applied current on the process yield, instantaneous current efficiency, overall current efficiency and specific energy consumption of the process. No difference in the powder production relative to the Base Case was observed. Page 4

54 Process Yield for Cu, XCu Process Efficiency (%) Process Yield for Copper A/m2 1 A/m2 15 A/m2 Overall Current Efficiency 14% 12% 1% 8% 6% 4% 2% % A/m2 1 A/m2 15 A/m2 Current Efficiency (%) Specfic Energy Consumption (kwh/kg Cu) Instantanious Current Efficiency 14% 12% 1% 8% 6% 4% 2% % A/m2 1 A/m2 15 A/m2 Specific Energy Consumption A/m2 1 A/m2 15 A/m2 Figure 34: Impact of applied current on various process parameters. Table 6 summarises the processing time, overall current efficiency, cell voltages and specific energy consumptions at various current densities and targeted copper extractions. Table 6: Processing parameters at various current densities and targeted Cu extractions. Targeted process yield (%) 9% Current density (A/m 2 ) Processing time (min) Overall current efficiency (%) Cell voltage (V) Spec. energy consumption (kwh/kg Cu) % % % % % % % % % % % From Table 6 it can be seen that for targeted process yields of 9%, 95% and 99% the processing time is 33-36% longer at 75 A/m 2 and 34-35% shorter at 1 5 A/m 2 (relative to Base Case). In general the specific energy consumption varies from kwh/kg Cu. The specific energy consumption together with required processing time for a targeted process yield of 95% is given in Figure 35 as a function of current density. Page 41

55 Specific Energy Consumption (kwh/kg Cu) Specific Energy Consumption vs. Processing Time Process Yield = 95% Current Density (A/m 2 ) Processing Spec. Energy Consumption Processing Time Figure 35: Specific energy consumption and processing times as a function of current density From Figure 35 it can be seen that for a targeted process yield of 95% the optimum current density is 1 A/m 2 in terms of lowest specific energy consumption and processing time. Based on this analysis it can be seen that higher applied current resulted in an increase in the rate of Cu recovery with a significant decrease in processing time. Conversely lower applied currents resulted in increased processing time, but generally have lower specific energy consumptions. For a targeted Cu recovery of 95% it was found that the optimum current density is 1 A/m 2 at a specific energy consumption of kwh/kg Cu Initial Acidity Based on the operating and recovery profiles for Runs 1A (Base Case), 4 (High Acidity) and 5 (Low Acidity) (Appendix B) it can be seen that the solution temperature and cell voltage increased with lower acidities. Copper, nickel and iron recovery profiles remained unchanged. At an initial acidity of.5 mol/l H + (Low Acidity) the electrolyte temperature increased to 68.1 C after 4.5 hours of processing (15.6 C greater than Base Case). This corresponded with higher cell voltages, which reduced from 9.15 V to 3.99 V after 4.5 hours of processing (.93 V higher than Base Case). At an initial acidity of 4 mol/l H + (High Acidity), the electrolyte temperature was 4.8 C lower than Base Case after 4.5 hours of processing. This corresponded with lower cell voltages from 3.6 V to 2.8 V after 4.5 hours of processing (.26 V lower than Base Case). Figure 36 illustrates the impact of initial acidity on the process yield, instantaneous current efficiency, overall current efficiency and specific energy consumption of the process. In the range of.5-4 mol/l H +, it was found that initial acidity of the aqueous solution did not significantly impact process yield, instantaneous current efficiency and overall current efficiency of the process. However, as a result of Page 42

56 the increased cell voltage at lower acidities, specific energy consumption was higher under these conditions. Process Yield Cu, XCu Process Efficiency (%) Process Yield for Copper mol/l 2 mol/l H+ 4 mol/l H+ Overall Current Efficiency 14% 12% 1% 8% 6% 4% 2% % mol/l H+ 2 mol/l H+ 2 mol/l Current Efficiency (%) Specfic Energy Consumption (kwh/kg Cu) Instantanious Current Efficiency 14% 12% 1% 8% 6% 4% 2% % mol/l H+ 2 mol/l H+ 4 mol/l H+ Specific Energy Consumption mol/l H+ 2 mol/l H+ 4 mol/l H+ Figure 36: Impact of initial acidity on various process parameters. The copper powder produced at an initial acidity of.5 mol/l H + (Low Acidity) and 5 mol/l H + (High Acidity) is shown in Figure 37a and Figure 37b, respectively. (a) (b) Figure 37: Copper powder produced under (a) Low Acidity and (b) High Acidity conditions. Page 43

57 From Figure 37a it can be seen that at the Low Acidity conditions the copper powder produced during the run was more adherent to the cathode surface (even after the electrolyte was drained and the cathode removed from the reactor. However, under High Acidity conditions (Figure 37b) similar copper powder production relative to the Base Case were observed. Thus, based on this analysis it is concluded that low initial acidities in the order of.5 mol/l H + result in elevated temperatures and copper powders that adhere more strongly to the cathode surface. Given the potential impact of elevated temperature on materials of construction, and the requirement for more rigorous solids removal from the cathode, it is recommended that initial acidities be in the order of 2-4 mol/l H Initial Copper Concentration Based on the operating and recovery profiles for Runs 1A (Base Case), 6 (High Cu) and 7 (Low Cu) (Appendix B) it can be seen that the solution temperature and cell voltage increased marginally with lower copper concentration. Slightly lower acidities were determined at lower copper concentrations. Nickel and iron recovery profiles remained unchanged. Figure 38 illustrates the impact of initial copper concentration on the process yield, instantaneous current efficiency, overall current efficiency and specific energy consumption of the process. Fractional Conversion of Cu, XCu Process Efficiency (%) Fractional Conversion of Copper g/l Cu 3 g/l Cu 54 g/l Cu Process Efficiency 14% 12% 1% 8% 6% 4% 2% % g/l Cu 3 g/l Cu 54 g/l Cu Current Efficiency (%) Specfic Energy Consumption (kwh/kg Cu) Current Efficiency 14% 12% 1% 8% 6% 4% 2% % g/l Cu 3 g/l Cu 54 g/l Cu Specific Energy Consumption g/l Cu 3 g/l Cu 54 g/l Cu Figure 38: Impact of initial copper concentration on various process parameters. From Figure 38 it can be seen that the initial copper concentration impacts the processing time, process efficiency and specific energy consumption. Lower copper concentrations required shorter processing times, but had lower process efficiencies and specific energy consumptions. The copper powder produced at an initial concentration of 3 g/l copper (Low Cu) and 54 g/l copper (High Cu) is shown in Figure 39 (a, c) and Figure 39 (b, d) respectively. Page 44

58 (a) (b) (c) (d) Figure 39: Pictures of the powders produced with (a,c) High Cu and (b,d) Low Copper runs. From Figure 39 it can be seen that at the Low Cu condition very fine black copper powder was produced. However, at the High Cu condition similar copper powder production relative to the Base Case was observed. Page 45